rajakhil.
← Back to all work

Python · AI · APIs · Workflow Engineering

Automation& Systems.

Custom tools that monitor information, identify what's relevant, and automate work that would otherwise require hours of manual effort.

Explore projects ↓Discuss an automation ↗

Featured System

Local AI & Lead Discovery

Discover relevant opportunities and prepare personalized responses without depending on paid cloud inference APIs.

Lead Discovery / Local AI

Reddit Opportunity Monitor

A custom monitoring system that finds relevant freelance hiring opportunities and prepares tailored application drafts using locally hosted AI.

PythonOllamaRedditTelegram APILocal LLM

The challenge

Searching multiple Reddit feeds, evaluating individual posts, and writing applications manually takes time and makes suitable opportunities easy to miss.

My approach

Python monitors configured feeds and filters incoming posts. Relevant opportunities are sent to Telegram alongside suggested replies generated by a local Ollama model. AI-assisted filtering can be enabled when needed.

System capabilities

  • Custom Reddit feed monitoring
  • Keyword-based and optional AI filtering
  • Context-aware application drafts
  • Telegram notifications
  • Configurable operating modes
  • Locally hosted Ollama inference

Delivered Output

Relevant hiring posts and tailored application drafts delivered to Telegram for review.

Local inference eliminates per-token cloud inference fees, although electricity and hardware still have costs.

Research & Publishing

Business Intelligence Pipelines

Automated information discovery, relevance ranking, document analysis, and curated publishing.

Business Intelligence

UK Insolvency Intelligence

An automated intelligence pipeline that processes UK Gazette insolvency notices and turns relevant events into curated newsletters.

PythonUK GazetteOpenAI APIRanking EngineNewsletter

The challenge

High volumes of public notices make it difficult to consistently identify and prioritize relevant companies and events.

My approach

A Python pipeline processes incoming notices, applies structured filtering and a custom ranking system, then uses the OpenAI API to help prepare editorial summaries.

System capabilities

  • UK Gazette monitoring
  • Company and notice extraction
  • Structured relevance filtering
  • Custom opportunity ranking
  • AI-assisted editorial summaries
  • Scheduled newsletter preparation

Delivered Output

Five curated newsletter editions per week.

Business Intelligence

US Insolvency Intelligence

A US-focused research automation system for identifying and prioritizing insolvency-related business information.

PythonPublic RecordsOpenAI APIRanking EngineNewsletter

The challenge

Insolvency information is distributed across multiple US data sources and requires careful filtering before it becomes useful.

My approach

The system collects information from configured US sources, normalizes relevant records, applies ranking logic, and produces AI-assisted newsletter material.

System capabilities

  • US-specific source collection
  • Record normalization
  • Event filtering and ranking
  • Automated research summaries
  • OpenAI-assisted content preparation
  • Recurring newsletter workflow

Delivered Output

Five curated newsletter editions per week.

Legal Document Intelligence

Court Docket Intelligence

A research automation system that reviews accessible high-profile court filings and identifies notable quotations for a weekly newsletter.

PythonCourtListenerDocument ProcessingText ExtractionNewsletter

The challenge

Potentially interesting public statements can be buried in extensive legal records, making manual discovery and verification difficult.

My approach

The pipeline uses CourtListener to examine relevant dockets and accessible documents, identify notable statements, and support source-aware editorial selection.

System capabilities

  • High-profile case monitoring
  • CourtListener integration
  • Accessible docket and document analysis
  • Quote identification and extraction
  • Attribution and source review
  • Weekly newsletter preparation

Delivered Output

One weekly newsletter featuring five selected quotations.

Statements are reviewed in context to distinguish direct quotations, allegations, testimony, and representations made in legal filings.

E-Commerce Automation

Systems Behind the Store

Practical integrations that reduce manual catalog management and keep store operators informed.

E-Commerce Operations

Automated Product Catalog Pipeline

A scalable WooCommerce production workflow for preparing and maintaining large product catalogs.

PythonWooCommerce APICanvaWebPBatch Processing

The challenge

Creating technical product listings with consistent descriptions, specifications, and multiple graphics becomes repetitive at scale.

My approach

Python workflows process structured product data, assist with content preparation, organize image assets, and update WooCommerce products through its REST API.

System capabilities

  • Batch product processing
  • Structured content preparation
  • Bulk graphic production workflows
  • WebP optimization
  • WooCommerce REST API updates
  • Auditing and recoverable batch operations

Delivered Output

A catalog workflow developed around approximately 100 products and hundreds of graphics.

Commerce Integrations

Telegram Commerce Notifications

A WooCommerce integration that delivers selected store events directly to Telegram.

WooCommerceWordPressPHPTelegram API

The challenge

Manually checking the store dashboard for new orders and operational updates interrupts regular work.

My approach

The integration transforms relevant WooCommerce events into readable Telegram notifications using custom event-handling and message-formatting logic.

System capabilities

  • WooCommerce event handling
  • Order activity notifications
  • Readable message formatting
  • Custom notification workflows
  • Integration into existing stores

Delivered Output

Relevant commerce events delivered as actionable Telegram notifications.

The right technology for the problem.

Not every task requires AI. Sometimes a carefully designed ranking algorithm, a scheduled script, or a simple API integration is the most reliable solution.

When language models are genuinely useful, I can integrate local Ollama models or hosted AI APIs, depending on the complexity, available hardware, and operating costs.

Explore automation services ↗

Have something worth automating?

From monitoring and reporting to data processing and custom integrations, let's discuss the requirements and an approach that makes sense.

Discuss your requirements ↗